A High-Precision Classification Method of Mammary Cancer Based on Improved DenseNet Driven by an Attention Mechanism

Table 6

Comparison of the results of binary classification experiments and multiclassification experiment with MFSCNet and other methods.

Methods	Classification task	Accuracy of different magnification (%)
		40x	100x	200x	400x
BiCNN	Binary classification	97.89	97.64	97.56	97.97
	Multiclass classification	—	—	—	—
CSDCNN	Binary classification	97.1	95.7	96.5	95.7
	Multiclass classification	94.1	93.2	94.7	93.5
CNN	Binary classification	98.33	97.12	97.85	96.15
	Multiclass classification	92.8	93.9	93.7	92.9
BHCNet	Binary classification	98.87	99.04	99.34	98.99
	Multiclass classification	93.74	93.81	92.22	90.66
MFSCNet A	Binary classification	99.51	99.46	99.89	99.05
	Multiclass classification	97.13	94.36	98.41	95.96